Start your memory layer. Yours for life.
- 3 personal boards
- 500 entries
- 1 connector key
- Text search
- No team boards
- No passive learning
- No semantic search
Free for personal use. Pro and Max unlock teams, passive learning, and unlimited boards. No usage taxes on the memory layer — once you've stored it, retrieval is free.
Start your memory layer. Yours for life.
For power users running 5+ AI tools.
For small teams that live in AI.
Limits and capabilities are defined in a single source of truth on the backend. What you see here is exactly what each tier is allowed to do.
| Capability | Free | Pro | Max |
|---|---|---|---|
| Storage | |||
| Personal boards | 3 | 50 | Unlimited |
| Team boards owned | 0 | 3 | 25 |
| Team members per board | 0 | 5 | 50 |
| Total entries | 500 | 25,000 | Unlimited |
| Connector keys | 1 | 10 | Unlimited |
| Capabilities | |||
| Passive learning | – | ✓ | ✓ |
| Semantic search | – | ✓ | ✓ |
| Operational | |||
| Audit log retention | 7 days | 90 days | 365 days |
| Connector requests / min | 30 | 300 | 3,000 |
| Proactive engine cadence | Daily | Hourly | Per write |
| Support tier | Community | Priority | |
A Board is a topic-scoped space in your memory layer — a place where a coherent set of knowledge lives. You might have a Music Production Board, a Business Board, a Travel Board.
An Entry is a single piece of structured memory inside a Board. Entries are written by you, by connected AI apps on your behalf, or by passive learning. The plan limits count Boards and Entries; everything else (reads, retrievals, embeddings) is unmetered.
Yes. Upgrades take effect immediately and are pro-rated against your current cycle. Downgrades take effect at the end of the cycle so you keep what you paid for. If a downgrade would put you over the new plan's limits, Boards and Entries stay intact — you just can't add new ones until you're back under cap or upgrade again.
No. The content of your Boards and Entries is never used to train AI models. The Orchestrator and Memory Engine call third-party LLM and embeddings APIs (currently Anthropic and OpenAI) under their no-training data usage policies; nothing is retained by them for training.
Passive learning is opt-in, off by default, and deletable at any time. See the Privacy Policy for the full breakdown.
Max supports up to 50 members on a team Board and 25 team Boards per workspace. If you need more — larger orgs, SSO, custom retention windows, or on-prem deployment — reach out at [email protected] and we'll talk.
Pro and Max are subscriptions billed monthly or annually through Stripe. You can cancel any time from the customer portal; service continues until the end of the period you paid for. Annual is roughly 15% cheaper than 12 months of monthly. No setup fees, no per-seat surcharges on Pro, no overage charges.
The INDEX backend is structured so that the orchestrator, memory engine, and connector layer can run on your own infrastructure. We don't publish a turnkey self-host package today, but enterprise self-host arrangements are available — get in touch and we'll walk through the deployment shape.